A simple process for the recovery of palladium from wastes of printed circuit boards

Journal Pre-proof A simple process for the recovery of palladium from wastes of printed circuit boards

Damien Bourgeois, Valentin Lacanau, Régis Mastretta, Christiane Contino-Pépin, Daniel Meyer PII:

S0304-386X(19)30822-9

DOI:

https://doi.org/10.1016/j.hydromet.2019.105241

Reference:

HYDROM 105241

To appear in:

Hydrometallurgy

Received date:

17 September 2019

Revised date:

5 December 2019

Accepted date:

19 December 2019

Please cite this article as: D. Bourgeois, V. Lacanau, R. Mastretta, et al., A simple process for the recovery of palladium from wastes of printed circuit boards, Hydrometallurgy(2019), https://doi.org/10.1016/j.hydromet.2019.105241

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2019 Published by Elsevier.

Journal Pre-proof

A simple process for the recovery of palladium from wastes of printed circuit boards

Damien Bourgeois a, Valentin Lacanaua,b, Régis Mastrettaa, Christiane

a

oo f

Contino-Pépinb, Daniel Meyera Institut de Chimie Séparative de Marcoule, ICSM, CEA, CNRS, ENSCM,

Equipe Chimie Bioorganique et Systèmes Amphiphiles, Institut des

e-

b

pr

Univ Montpellier, BP 17171, Marcoule, 30207 Bagnols-sur-Cèze, France.

Pr

Biomolécules Max Mousseron, UMR 5247, Avignon Université, 84911

rn

al

Avignon, France.

Jo u

[email protected]

Journal Pre-proof

ABSTRACT: An efficient process for the recovery of palladium from waste printed circuits

2

boards (PCBs) is detailed. Palladium is employed as an electrode material in multi-layer

3

ceramic capacitors (MLCCs). These components can be easily removed from PCBs by de-

4

soldering. As palladium is alloyed with silver, its dissolution is readily achieved using dilute

5

nitric acid. As a result, a solution containing palladium along with base metals, mostly

6

copper and iron, is obtained. This solution is then processed through solvent extraction

7

(SX) with a solvent based on N,N’-dimethyl,N,N’-dibutyltetradecylmalonamide (BDMA), a

8

robust extracting molecule previously developed in the frame of the reprocessing of waste

9

nuclear fuel. The volume of effluents generated during the SX sequence is limited: iron

10

scrubbing is operated with a very low aqueous to organic phase volume ratio, no specific

11

metal chelator is required for palladium stripping, and no shift from acidic to basic media is

12

required. Finally, a ca 1 g/L Pd(II) aqueous solution with 99,4% purity is obtained, from

13

which palladium is directly isolated as dichlorodiammine palladium(II) salt (Pd(NH3)2Cl2)

14

after precipitation with ammonia. Overall, palladium is quantitatively recovered from the

15

leaching solution, and no palladium was detected in the remaining solid residue. Purity is

16

high, as no contaminating metal could be detected in the final palladium salt. The

17

proposed approach is simple and complementary to existing hydrometallurgical processes

18

dedicated to gold and copper recovery.

Jo u

rn

al

Pr

e-

pr

oo f

1

19 20

Keywords:

21

Palladium; e-waste; Solvent extraction; Malonamide

Journal Pre-proof

1. Introduction

23

Palladium is a precious metal which is principally employed in autocatalysts. The

24

increasing demand in palladium for this application combined with stable production has

25

recently led to a sharp rise in palladium price as the market remains tight (Cowley, 2019).

26

Palladium price more than doubled over the last three years, and even, in January 2019,

27

reached that of gold. Since then, palladium price regularly reaches ca. 45 k€/kg

28

(1600 $/oz). Palladium market is expected to go further in deficit and production from other

29

secondary resources

30

Ragnarsdottir, 2016). Supply from the so-called urban mine appears as a very attractive

31

alternative to traditional mining, especially for Western countries, poorly endowed with

32

natural mineral ores (Sun et al., 2016). Collected waste volume is increasing, and largely

33

unexploited resources are regularly disclosed, such as printed circuits boards (PCBs)

34

contained in end of life vehicles (Xu et al., 2019). Gold represents the principal value

35

fraction of metals contained in waste of electric and electronic equipment (WEEE) (Diaz et

36

al., 2016). Copper and silver represent respectively the highest base-metal and semi-

37

precious metal contents and are present in important quantity. Processes dedicated to the

38

recovery of metals from WEEE have been widely investigated, and mostly optimized for

39

gold and copper recovery (Hsu et al., 2019; Işıldar et al., 2018; Kaya, 2018). Precious

40

metals apart from gold have been generally much less detailed in these studies (Lu and

41

Xu, 2016). Existing copper smelters have been successfully adapted for the reprocessing

42

of WEEE, and several industrial companies claim to recover palladium along with gold and

43

copper from WEEE (Cui and Zhang, 2008). This technical solution is well-adapted to the

44

reprocessing of used autocatalysts, as precious metals are supported on an inert alumino-

45

silicate (cordierite) (Ding et al., 2019). But reprocessing of used PCBs in copper smelters

a

key to

fulfil palladium needs

(Sverdrup

and

Jo u

rn

al

Pr

e-

pr

becomes

oo f

22

1

Journal Pre-proof necessitates complicated off-gas treatments (Cui and Zhang, 2008; Li et al., 2018). In

47

complement, hydrometallurgical processes have been developed to recover valuable

48

metals from waste PCBs: These are generally soaked in acidic baths to remove unwanted

49

base metals (copper, iron, nickel…), which leaves a solid residue, enriched in gold, further

50

processed through dissolution/isolation sequences (Birloaga and Vegliò, 2018). Very often,

51

the outcome of palladium in these processes is not detailed. Efficient techniques targeting

52

palladium recovery from secondary resources are scarce to our knowledge. Detailed

53

analysis of the proposed approaches reveals the difficulty to design a simple process

54

which enables isolation of pure palladium (Table 1). All approaches rely on poorly selective

55

halide based precious metals leaching agents (chloride or iodide media), and (when

56

operated) the purification of palladium can be tedious. Proposed processes often end up

57

with a mixture, of low or unspecified palladium concentration, and isolation of pure

58

palladium is seldom performed. Finally, some proposed processes clearly raise safety

59

issues (Table 1). As 8 to 9% of the total palladium demand is employed in electronic goods

60

manufacturing, it is clearly understandable that palladium recycling from waste PCBs

61

deserves a much deeper consideration.

62

Available data suggest that palladium concentration in waste PCBs ranges between 10

63

and 100 mg/kg (Yazici and Deveci, 2013; Zhang and Zhang, 2014). Higher concentrations

64

have been found in specific cases, though on quite old samples (Cui and Zhang, 2008). In

65

PCBs, palladium is principally employed in multi-layer ceramic capacitors (MLCCs)

66

because of its electrical conductivity and its durability (Işıldar et al., 2018; Prabaharan et

67

al., 2016). Palladium is found in the conductive electrode material sandwiched between

68

insulating ceramic layers. Apart from very old MLCCs (more than 20 years old), where

69

100% Pd electrode material was employed, palladium is found in an alloy with silver (Lee,

70

2010; Wang et al., 1994). A cheaper alternative to the use of silver-palladium alloys is the

Jo u

rn

al

Pr

e-

pr

oo f

46

2

Journal Pre-proof use of nickel electrodes, which leads to heavier and less durable MLCCs. As nickel is a

72

ferromagnetic metal, different MLCC classes can be easily discriminated with a simple

73

magnet: Base metal electrodes (BME) MLCCs are magnetic, whereas precious metal

74

electrodes (PME) MLCCs are not. Most of precious metals recovery processes from waste

75

PCBs start with a shredding or grinding step (Table 1). The purpose of this step is to

76

ensure access to the metallic parts for the leaching/dissolving agent, and is directly

77

inspired from the comminution step of natural ores processing. In the case of waste PCBs,

78

a different approach is also sometimes considered as i. metallic parts are often easily

79

accessible (surface layers, continuous contact…) and ii. reducing PCBs into powder also

80

leads to total mixing of the metals. Several studies focusing on gold and copper suggest

81

removing electronic components from boards prior to leaching in order to lower the amount

82

of leaching agent required (Choudhary et al., 2017; Fontana et al., 2018). Components are

83

connected to the boards through solders, mostly made of Sn-Pb based alloys in waste

84

PCBs, which can easily be removed either via thermal or chemical treatment (Kaya, 2018).

85

This simple step enables obtaining electronic components separated from naked boards.

86

The location of palladium in MLCCs pushes towards such an approach, which enables

87

simple separation of palladium and gold/copper flows.

88

In this work, we describe a complete detailed study which enables the recovery of

89

palladium from used PCBs, through a hydrometallurgical process based on chemical de-

90

soldering of components, including MLCCs, followed by selective leaching with nitric acid,

91

selective

92

dibutyltetradecylmalonamide, BDMA, Figure 1) based solvent, and precipitation of a pure

93

palladium salt using standard techniques. The choice of the extracting molecule was

94

motivated by our previous results which demonstrated its excellent performance in the

95

processing of complex aqueous nitrate solutions containing palladium (Mastretta et al.,

Jo u

rn

al

Pr

e-

pr

oo f

71

3

palladium

extraction

with

a

malonamide

(N,N’-dimethyl,N,N’-

Journal Pre-proof 96

2019; Poirot et al., 2016), along with the consequent technical knowledge gathered during

97

process development studies dedicated to spent nuclear fuel reprocessing with same

98

molecule (Modolo et al., 2007). The adaptation of an existing technology for a new

99

application thus only deserves proof of concept at the laboratory scale, as most process

100

issues have already been tackled out. Overall, the proposed process generates a solid

101

residue which contains most initial gold and copper content, and a controlled volume of

102

aqueous effluents, which can be in part reemployed in the leaching step.

oo f

103

pr

104

2. Experimental

106

2.1 Materials and reagents

107

Used electronic devices were picked up randomly from WEEE collection points and

108

manually dismantled to recover the PCBs contained inside. Concentrated nitric acid,

109

concentrated hydrochloric acid, sodium chloride, toluene and aqueous 28% ammonia

110

solution

111

dibutyltetradecylmalonamide (BDMA, Figure 1) was kindly provided by the Commissariat à

112

l’Energie Atomique et aux Energies Alternatives (CEA, France).

purchased

Jo u

were

rn

al

Pr

e-

105

from

Carlo

Erba

reagents.

N,N’-dimethyl,N,N’-

113 114

2.2 Dissolution of solders

115

A 3 M aqueous nitric acid solution (200 mL) was placed in a beaker and heated at 50°C.

116

Entire PCBs were placed in the solution and regularly gently swirled into the solution until

117

components fall to the bottom of the beaker (maximum 15 min). The boards were

118

removed, and the components recovered after filtration. The solution became turbid, and a

119

white solid settled in the bottom, and it was isolated after centrifugation. 4

Journal Pre-proof 120 2.3 Complete digestion of components

122

With aqua regia: Components were weighed and placed in a glass vessel. A mixture of

123

32% HCl and 69% HNO3 2/1 (v/v) at a 25 g/L solid/liquid ratio (S/L = 1/40) was then

124

added, and the vessel was covered with a glass slip. The resulting mixture was stirred at

125

80°C for 24 h using an IKA-Mag magnetic stirrer and a teflon coated bar. The resulting

126

suspension was filtered with a 0.20µm filter and the resulting solution analyzed by

127

ICP/AES. The residual solid was analyzed by SEM.

128

With aqueous nitric acid solution: The same procedure was applied, using a 3 M aqueous

129

HNO3 solution.

130

e-

pr

oo f

121

2.4 Palladium leaching from MLCCs

132

Components were weighed and placed in a glass beaker. A 3 M aqueous HNO 3 solution at

133

a 100 g/L solid/liquid ratio (S/L = 1/10) was then added, and the beaker was covered with

134

a glass slip. The resulting mixture was stirred at 80°C for 4 h using an IKA-Mag magnetic

135

stirrer and a teflon coated bar. The resulting suspension was centrifuged to remove fine

136

particles (ceramics, remaining tin(IV) oxide).

Jo u

rn

al

Pr

131

137 138 139

2.5 Solvent Extraction

140

Extractions were carried out in 50 mL glass vials by putting in contact the aqueous solution

141

resulting from leaching of components with an organic phase containing BDMA in toluene.

142

The two phases were vigorously shaken at 1000 vibrations per min with an IKA-Vibrax

143

VXR basic shaker in a 20-22°C environment during 1 h. The vials were removed and let to

144

stand for 5 min and the phases were separated. Scrubbing and stripping stages were

5

Journal Pre-proof 145

performed in 50 mL glass vials by putting in contact the organic phase resulting from

146

previous stage, with an adequate aqueous solution (diluted nitric acid solution, pure water

147

of sodium chloride solution, see text). The volume of the aqueous phase was chosen

148

according to the A/O ratio detailed in the text. The two phases were vigorously shaken at

149

1000 vibrations per min with an IKA-Vibrax VXR basic shaker in a 20-22°C environment

150

for 1 h. The vials were removed and let to stand for 5 min and the phases were separated.

151

When necessary, 500 µL of each phase were taken for analysis.

oo f

152 2.6 Palladium precipitation

154

To the aqueous sodium chloride solution containing Pd(II) was added dropwise an

155

aqueous 28% ammonia solution until the solution became clear and colorless and the

156

medium became basic (ca. 5% of the initial volume needed). To the resulting solution, a

157

6 M aqueous hydrochloric solution was added dropwise under constant stirring until

158

precipitation of a yellow solid. The resulting suspension was stirred until it cooled to room

159

temperature, and filtered. The resulting solid was washed with water and dried in an oven

160

at 40°C overnight.

Jo u

rn

al

Pr

e-

pr

153

161 162

2.7 Determination of metal concentration in solutions

163

Aliquots of aqueous phases were directly diluted into a 2% HNO 3 aqueous solution. Metals

164

contained in the aliquots of the organic phases were back-extracted with 800 µL of an

165

aqueous 0.1 M thiourea solution at 20-22°C for 1 h. Both phases were separated, and

166

500 µL of the resulting aqueous solution were taken and diluted into a 2% HNO 3 aqueous

167

solution. Concentrations of each metal in both aqueous phases (extraction and stripping

168

phases) were determined by inductively coupled plasma atomic emission spectroscopy

169

(ICP/AES, SPECTRO ARCOS ICP Spectrometer, AMETEK Materials Analysis). Given

6

Journal Pre-proof 170

concentrations are calculated as the means of three replicates on three different

171

wavelengths for each metal; relative standard deviations were determined and lie between

172

1 and 4%. Quantification limits (LoQ) in each analyzed phase were determined for each

173

metal from the dilution factor applied and from the background equivalent concentration

174

(BEC), calculated by the spectrometer for each optical line.

175 2.8 Powder X-ray diffraction (PXRD)

177

The PXRD pattern of the obtained solid was recorded at ambient temperature on a Bruker

178

D8 ADVANCE powder diffractometer using Ni filtered CuK  (=1.5406 Å) radiation. Data

179

was collected in the range 5-60° (2θ), with a step of 0.01°, and a scanning rate of 0.1°/min.

pr

oo f

176

e-

180

2.9 Scanning Electronic Microscopy (SEM) and elemental analysis

182

SEM characterization was obtained with a FEI QUANTA FEG 200 environmental scanning

183

electron microscope (ESEM) equipped with a gaseous secondary electron detector

184

(GSED). Experiments were performed at 100 Pa and an accelerating voltage of 15 kV.

185

Elemental analysis was performed with X-ray Energy dispersive spectrometry (X-EDS)

186

analyses using a Bruker AXS X-flash 5010 detector coupled to the ESEM.

Jo u

rn

al

Pr

181

187 188

189

3. Results and discussion

190

3.1 Chemical de-soldering and isolation of MLCCs

191

Thermal de-soldering of components from PCBs has been implemented at the industrial

192

scale. Its application requires specific equipment and temperature control to enable

193

complete melting without emission of toxic lead fumes. On smaller scale, it may be 7

Journal Pre-proof interesting to turn to chemical de-soldering (Zhang et al., 2015). Several media have been

195

employed, and the most efficient technique appears to be immersion into an aqueous nitric

196

acid bath (Yang et al., 2011; Yoo et al., 2012). We evaluated at laboratory scale the

197

removal of electronic components from boards after immersion into 3 M aqueous nitric

198

acid solution. Small pieces (ca. 2 x 2 cm2) of different PCBs were cut off and immersed

199

into 3 M nitric acid. According to the board studied, complete dissolution was either

200

observed after a few minutes at room temperature, or after 10 to 15 min of gentle heating.

201

This behavior may be related to the composition of solders employed and/or the presence

202

of varnish. Altogether, a 3 M aqueous nitric acid bath was prepared and heated at 50°C in

203

order to achieve removal of components from all PCBs studied. Most small components

204

fall to the bottom of the bath, while some pinned components remain on the board.

205

Interestingly, MLCCs are not pinned on the boards (Figure 2), and pinned components are

206

generally memory bars or processors, which often contain gold. De-soldered components

207

are then recovered after simple filtration.

208

Chemical de-soldering has several advantages. First of all, it is a safe operation. Only

209

dilute aqueous nitric acid is employed, no brown NOx fume could be observed. It is a rapid

210

process, as tin, lead and Sn-Pb alloys dissolutions are very easy in nitric acid. Secondly,

211

no gold is leached into nitric acid in the absence of chloride ions, and copper is almost not

212

attacked during this process. The de-soldering solution was analyzed (Table 2) and

213

contains Cu at a g/L level, but this does not represent an important quantity considering

214

the large amount of Cu contained in PCBs. A rough estimate based on an average weight

215

Cu content of 15% in PCBs suggests that more than 95% of Cu remains on the board.

216

Thirdly, tin can be easily recovered in the process. A white precipitate was observed in the

217

filtrate after few minutes, and its isolation followed by X-EDS analysis revealed that it is

218

exclusively composed of Sn. Actually, oxidation of metallic tin by nitric acid has been

Jo u

rn

al

Pr

e-

pr

oo f

194

8

Journal Pre-proof described to lead to complete oxidation into Sn(IV) species, which are soluble only in the

220

presence of chloride ions. Otherwise, hydrolysis of the metallic cation occurs, and leads

221

eventually to precipitation of Sn(IV) oxide (as SnO2 or H2SnO3), as long as concentrated

222

nitric acid is not employed (Yoo et al., 2012). Finally, no Ag nor Pd were detected in the

223

solution, so that it is relatively uncomplicated, and contains only Cu, Pb, Ni and Fe as

224

major metallic salts (Table 2). Re-use and reprocessing of such solutions have been

225

already successfully investigated by several authors (Lee et al., 2003; Yang et al., 2017).

226

As a consequence, we have not fully optimized this step.

227

Components recovered after filtration are of various natures. It is possible to operate

228

following leaching step on the whole mixture (Figure 2), or to remove magnetic

229

components and/or processors (black components) and plastic parts. It is also possible to

230

isolate large non-magnetic MLCCs to check their integrity and evaluate their Pd content. At

231

laboratory scale, all these steps were performed manually. But physical techniques to

232

operate sorting of WEEEs pieces have been well developed and can easily be envisioned.

al

Pr

e-

pr

oo f

219

rn

233

3.2 Location of Pd in waste PCBs and consequences for leaching

235

Several non-magnetic MLCCs were selected and collected after de-soldering, and

236

analyzed with optical and scanning electronic microscopy (SEM). The multilayer structure

237

is well visible at low magnification with optical microscopy, but the electrode can only be

238

seen thanks to SEM as its thickness is only in order of 1 µm (Figure 2). Chemical analysis

239

by X-ray Energy dispersive spectrometry (X-EDS) revealed the presence of both Ag and

240

Pd, in a ratio varying between 4/1 and 2/1 according to the considered sample. Pd was

241

always found along with silver, confirming the use of Ag-Pd alloys. Barium and titanium

242

were the main elements found in the thick layer, in full agreement with the well-described

243

use of a BaTiO3 ceramic as the dielectric material. Ag was also detected in the connecting

Jo u

234

9

Journal Pre-proof terminations of the MLCCs, pure in most cases, containing a small amount of Pd in some

245

cases (up to 5% respective to Ag). Similar analysis performed on magnetic MLCCs

246

revealed the use of Ni or Ni-Sn based electrodes, with Ni or Cu based connecting

247

terminations.

248

In order to evaluate the amount of palladium present in MLCCs, we tried to perform

249

complete digestion with aqua regia of one MLCC reduced to coarse powder after manual

250

grinding in a mortar. Almost no Pd was detected in the leaching solution (4 mg/L). A white

251

solid residue was left, and its analysis by X-EDS revealed the presence of both Ag and Pd

252

in the solid, along with Ti from the ceramic. This result is quite surprising since aqua regia

253

is often proposed for complete dissolution of precious metals present in waste PCBs.

254

Previous studies give however contradictory information about the leaching of Pd and Ag

255

by aqua regia: Metallic palladium leads to soluble chloro complexes in aqua regia even in

256

the presence of Ag (Jung et al., 2009; Prabaharan et al., 2016). The reduction of nitric acid

257

into NH4+ could lead to the precipitation of stable red Pd(NH4)2Cl6 without the need for

258

stronger oxidizing agent to stabilize the Pd(IV) in the system. This has been observed in

259

the presence of Zn(II) or Ni(II) ions (Park and Fray, 2009). Ni(II) has been quantified in our

260

study (550-1100 mg/L), but no red precipitate was observed. Silver is not soluble in dilute

261

chloride media, but can be solubilized by concentrated aqua regia as AgCl2-, and this has

262

been observed also using a low 2/1 (v/v) ratio of hydrochloric to nitric concentrated acids

263

(Wang et al., 2016). Surface passivation by AgCl has also been described during attack of

264

metallic Ag alone in aqua regia even at low concentration, using a 1/40 S/L ratio (Park and

265

Fray, 2009). It is thus reasonable to assume that no dissolution of the Ag-Pd alloy

266

occurred, and that Pd dissolution was prevented by the presence of Ag. Two hypotheses

267

can be put forward: the temperature may be too low, as sometimes complete digestion is

268

performed at higher temperature in pressure vessels, or the S/L ratio may be still too high.

Jo u

rn

al

Pr

e-

pr

oo f

244

10

Journal Pre-proof We therefore turned to the use of nitric acid alone. Ag is readily soluble in nitric acid and 3

270

to 6 M aqueous nitric acid solutions have been successfully employed for efficient Ag

271

leaching from metals mixtures (Petter et al., 2014). Dissolution of Pd in the absence of

272

chloride ions may be sluggish, but its oxidation is accelerated through alloying with Ag

273

(Cole, 1985). Although available literature regarding selective leaching of palladium alloys,

274

especially Ag-Pd alloys, is scarce, we were pleased to find that dissolution using 3 M

275

aqueous nitric acid instead of aqua regia led to efficient Pd dissolution, even after only 4 h.

276

Although it is not possible to state whether Pd leaching is complete, no Pd was detected in

277

the remaining solid, and Pd content in the MLCCs studied was determined to be between

278

0,5 and 3% wt (5 to 30 g/kg). Latter value demonstrates that Pd content in MLCCs can be

279

very high, and that focusing a recovery process on MLCCs should enable to recover

280

substantial Pd contained in waste PCBs. Finally, these results demonstrate that the use of

281

aqua regia for the dissolution of complex matrices containing precious metals is not so

282

simple. It would be possible, using a higher temperature or a lower S/L ratio (i.e. a higher

283

volume of liquid), but this approach does not appear as a good process solution.

284

It is conceivable to use the same nitric acid solution to perform first chemical dissolution of

285

solders then leaching of Pd from components. But such a strategy leads to a solution

286

rather diluted in Pd (3 to 5 mg/L). It appears thus judicious to employ a first bath for de-

287

soldering, regenerate it according to existing techniques, and then to perform leaching of

288

Pd from components with another aqueous solution, of limited volume. It is also possible to

289

select from the mixture of components obtained after de-soldering only non-magnetic

290

MLCCs. Preliminary leaching experiments (not optimized) led to Pd containing solutions of

291

relatively simple composition (Table 2). On the other hand, dissolution of magnetic

292

components led to mixtures containing no Pd, but relatively important concentration of Cu,

293

Fe and Ni instead (Table 2). As in both case contaminating metals are the same, and only

Jo u

rn

al

Pr

e-

pr

oo f

269

11

Journal Pre-proof their concentrations differ, the selection of specific MLCCs from the components mixture is

295

not necessary. Indeed, leaching of metals from a mixture of components not selected

296

(Figure 2) led to a solution containing ca. 100 mg/L Pd along with Ag, Cu, Fe, Ni and Pb at

297

higher concentration, and low amount of Sn and Zn (Table 2).

298

3.3 Separation and purification of Pd from a nitric acid solution

299

Palladium recovery from nitrate media resulting from the dissolution of used nuclear fuel in

300

nitric acid has been largely documented (Kolarik and Renard, 2003). However, we recently

301

highlighted that the proposed solutions are not adequate for a mixture containing common

302

base metals such as Fe, Cu, Pb, Zn, Ni, etc…, and turned to the use of malonamide based

303

solvents (Mastretta et al., 2019). Malonamides make up a family of extracting molecules

304

developed in the frame of the reprocessing of spent nuclear fuel (Musikas, 1988). Although

305

dedicated to the actinide(III)/lanthanide(III) separation in advanced purification process,

306

e.g. DIAMEX (Facchini et al., 2000; Sypula et al., 2012), their affinity for Pd(II) was also

307

evidenced (Poirot et al., 2014). The optimal process in the nuclear fuel cycle is based on

308

N,N’-dimethyl,N,N’-dioctylhexylethoxymalonamide (OEMA, Figure 1), but the previous

309

candidate, BDMA, showed a higher affinity for Pd(II) (Mastretta et al., 2019; Serrano-

310

Purroy et al., 2005). Latter molecule is also simpler to prepare, as the tetradecyl chain is

311

commercially available. Both molecules are very robust, stable after prolonged contact

312

with concentrated nitric acid, and even resistant to radiolysis (Le Caër et al., 2012).

313

Process studies performed with continuous feeds in mixer-settlers revealed nice phase

314

splitting (Modolo et al., 2007), probably thanks to low viscosity of the solvent (less than

315

10 cP). Benchmarking with commercially available extractants recently demonstrated that

316

malonamides are very well positioned and should be considered for the recovery of

317

palladium from complex mixtures such as those arising from the dissolution of electronic

318

waste (Mastretta et al., 2019).

Jo u

rn

al

Pr

e-

pr

oo f

294

12

Journal Pre-proof Regarding the present study, two key points need to be addressed in order to validate the

320

malonamide technology in Pd recovery processes: i. the possibility to generate low amount

321

of effluents during solvent extraction step, and ii. the effective isolation of Pd from waste

322

PCBs at laboratory scale. Considering the composition of the leaching solution obtained

323

beforehand, we evaluated the impact of each metal separately on the extraction step

324

(Table 3). It is clear that Fe is the impurity of concern, and that Cu should be also

325

controlled. All other metals have a lower distribution coefficient and/or a lower

326

concentration in aqueous phase, so that their concentration in organic phase is lower, and

327

that scrubbing of both Fe and Cu will guarantee their elimination from organic phase.

328

Suitable Pd distribution coefficients and Pd/Fe selectivity have been achieved using BDMA

329

in toluene (Poirot et al., 2016). Modelling a counter current process based on Pd extraction

330

step, Fe scrubbing step and Pd stripping step, each of one with 6 stages, leads to

331

quantitative Pd recovery (Mastretta et al., 2019). As it is much simpler to operate cross-

332

current processes at laboratory and pilot scale, we evaluated a cross-current process at

333

the laboratory scale. The key step in the Pd isolation sequence is the Fe scrubbing step,

334

where minimal Pd loss has to be achieved, along with complete Fe scrubbing generating

335

minimum amount of aqueous effluents. One of the interesting features of the malonamide

336

technology is that no extra scrubbing or stripping agent is required: only careful control of

337

the HNO3 concentration in the aqueous phase is required. The distribution coefficients

338

(D M) of both Pd and Fe directly depend on aqueous HNO3 concentration, and as long as

339

final aqueous HNO3 concentration is kept above 1,5 M, Pd distribution coefficient is well

340

above 1, and Pd/Fe selectivity well above 100 (Table 4). This enables quantitative Pd

341

extraction, along with residual Fe, as this metal can be in much higher concentration in the

342

initial aqueous phase than Pd. This residual Fe has to be scrubbed from organic phase to

343

ensure recovery of pure Pd. The concentrations of metals back-extracted from the organic

Jo u

rn

al

Pr

e-

pr

oo f

319

13

Journal Pre-proof 344

phase into the aqueous phase depend on distribution coefficients, along with the organic

345

to aqueous (O/A) phase volume ratio, according to equation (1): [ ]

[ ]

346

⁄

(1)

⁄

347 348

As a consequence, for a metal with a high distribution coefficient such as Pd, its loss

349

during a scrubbing step will directly depend on O/A and D Pd, following equation (2): [

] ]

⁄

⁄

(2)

oo f

[

350 351

And for a metal with a very low distribution coefficient such as Fe, its scrubbing from

353

organic phase will be almost quantitative, as long as its concentration in aqueous phase

354

does not exceed saturation. Its final concentration in aqueous phase can be approached

355

using equation (3): [

] ⁄

⁄

[

]

⁄

(3)

rn

357

]

al

[

356

Pr

e-

pr

352

From a practical point of view, there is a compromise to find between the O/A ratio and the

359

nitric acid concentration of the aqueous phase employed for scrubbing. A high O/A ratio

360

will enable a low Pd loss, along with low generation of effluents. A low O/A ratio will enable

361

an optimal Fe scrubbing. A low aqueous HNO3 concentration will enable efficient Fe

362

scrubbing but will also increase Pd loss. In a first instance, we evaluated both Fe

363

scrubbing (elimination) and Pd loss after contacting an organic phase loaded with both

364

metals with an aqueous phase at different O/A ratios (Figure 3). Using only water leads to

365

almost quantitative Fe scrubbing, albeit with important Pd loss (up to 5% at a O/A ratio of

366

5). Using 1,5 M HNO3 leads to lower Pd loss (less than 2% when O/A ratio exceeds 7), but

367

also lower Fe elimination (70 to 80% only). Altogether it appears clearly that at least two

Jo u

358

14

Journal Pre-proof stages have to be set in place, and different strategies based on sequential two stage

369

scrubbing were evaluated (Figure 4). The best option consists in a first scrubbing with

370

water only, followed by a 1 M aqueous HNO3 scrubbing, both with an O/A ratio of 10. In

371

these conditions Pd purity is 96% with 3% Pd loss. These screening results demonstrate

372

that an efficient Fe scrubbing step can be obtained with minimum generation of effluents:

373

the total volume of waste aqueous phase generated represents only 20% of the volume of

374

the organic phase. The final organic phase containing Pd can then be processed in order

375

to recover Pd as a Pd(II) salt with classical techniques.

oo f

368

376

3.4 Impact of chlorides and consequences for Pd recovery

378

Evaluation of the behavior of each metal in the solvent extraction step drove us also to

379

evaluate the impact of the presence of chloride ions in the aqueous phase. Sn is poorly

380

soluble in pure nitrate media, as Sn(IV) precipitates as mentioned beforehand. As a

381

consequence it is difficult to evaluate its distribution coefficient with precision as

382

concentration of Sn in the organic phase after extraction is below quantification limit. On

383

the other hand, similar experiment performed with a Sn(IV) chloride solution leads to

384

almost quantitative extraction of Sn. The influence of chloride ions was evaluated on 4 key

385

metals (Pd, Ag, Fe and Sn, Table 5). It is clear that chloride have a negative impact on the

386

process: i. they inhibit the leaching of Pd (see above) and lead to the precipitation of Ag, ii.

387

they decrease the efficiency of Pd extraction, and the Pd/Fe selectivity as Fe is extracted

388

more easily into organic phase, iii. they increase the solubility of Sn and the subsequent

389

extraction of Sn in the organic phase. Altogether, the use of aqua regia is not only poorly

390

suitable for leaching, but also for Pd recovery. Of course other extraction agents could be

391

envisioned (Fontana et al., 2018), but Sn will require deeper studies as it is employed in

392

the extraction of platinum group metals to increase their extraction efficiency: Sn(II)

Jo u

rn

al

Pr

e-

pr

377

15

Journal Pre-proof chloride is a labilizing agent which is easily co-extracted with Pd or Pt in chloride media

394

(Mojski, 1980). Furthermore, the use of a chloride solution to strip Pd from organic phase

395

appears as a very interesting option for malonamides: With low distribution coefficient in

396

the presence of chloride no specific Pd complexing agent will be required to strip Pd, from

397

organic phase. Afterwards, Pd(II) salts can be easily precipitated from aqueous chloride

398

solutions as Pd(NH3)2Cl2 salt through classical process.

399

Altogether, the complete sequence is the following: i. chemical de-soldering, ii. Pd leaching

400

from components, iii. Pd isolation through solvent extraction, iv. Pd precipitation as

401

Pd(NH3)2Cl2. The two last points are detailed in Figure 5 and Figure 6. A selected solution

402

containing a high amount of Fe and Cu (along with other metals Ag, Ni, Pb, Sn, Zn and

403

traces Al, Ba, Nd, Ti and Y, discarded for clarity) was employed and led to an aqueous

404

solution containing 85% of the initial Pd amount with a 99,4% purity: traces of Fe were still

405

quantified in this final Pd solution, Cu was below quantification limit, and other metals were

406

not detected (Figure 6). This aqueous solution was further processed through successive

407

addition of an ammonia solution and hydrochloric acid solution to lead to quantitative

408

precipitation of Pd(NH3)2Cl2 of characteristic yellow color. SEM, X-EDS and powder X-ray

409

diffraction (PXRD) analysis confirmed the structure and the purity of the salt (Figure 5). No

410

residual Fe was detected in the solid.

Jo u

rn

al

Pr

e-

pr

oo f

393

411 412

3.5 Regeneration of solvent

413

We chose not to fully back-extract (strip) Pd from the organic phase in order to minimize

414

the volume of aqueous phase employed: 15% of the Pd remains in the organic phase.

415

Theoretically, a 15% Pd hold-up in the extraction solvent should lead to a maximum 17%

416

(1 – 1 / 0,85) increase in the Pd loading in the organic phase after re-use of the organic

417

phase for a new solvent-extraction cycle. Such a slight increase in Pd concentration

16

Journal Pre-proof should not modify distribution coefficient nor extraction yield. Actually, re-use of this

419

solution for a new extraction cycle led to quantitative Pd extraction from same initial

420

aqueous phase. Organic phase Pd content is higher, and those of Fe and Cu do not vary

421

significantly. Interestingly, no regeneration of the organic phase was required: usually,

422

stripping the metal of interest with a strong chelator requires washing of the organic phase

423

in order to perform a new extraction cycle. In the present case, possible traces of chloride

424

did not alter Pd extraction. No other aqueous effluents are thus generated. Altogether,

425

starting from one volume of aqueous phase, the same volume of organic phase is needed

426

and can be re-employed in successive cycles, and effluents generated account for only

427

30% of the initial volume to be treated.

428

The choice of solvent is the direct consequence of our former studies on such systems

429

(Mastretta et al., 2019; Poirot et al., 2014). Using BDMA in toluene, we were able to tackle

430

key issues related to the development of a solvent extraction process, i.e. low generation

431

of effluents and solvent regeneration. Working on the worst possible case, i.e. a leaching

432

solution of high Fe concentration, we obtained very satisfactory results. Our previous

433

results demonstrate that the use of aromatic diluent is only motivated by the optimization

434

of Pd/Fe selectivity (Poirot et al., 2016), and that otherwise Pd distribution coefficient and

435

solubility in the diluent is marginally affected by shift to an aliphatic diluent (Poirot et al.,

436

2014). As toluene is volatile, it would be much safer to use a higher boiling point diluent,

437

such as an aromatic cut (e.g. Solvesso 150). Mixed aliphatic-aromatic cuts could also be

438

employed, and it would also be interesting to investigate malonamide in purely aliphatic

439

diluents when physical sorting of the waste guarantees low Fe content in the leachate. Any

440

of these solutions would also definitely limit inhalation exposure to the toluene, a harmful

441

substance.

Jo u

rn

al

Pr

e-

pr

oo f

418

442

17

Journal Pre-proof 443 444 445 446

5. Conclusion

448

Through this study, we demonstrated that efficient Pd recovery from waste PCBs can be

449

performed at laboratory scale via an approach based on successive de-soldering, Pd

450

leaching with dilute nitric acid, solvent extraction then precipitation. This approach takes

451

profit of the quasi-exclusive location of Pd in MLCCs, and the fact that it is alloyed with Ag.

452

It demonstrates that the systematic recourse to shredding and aqua regia is inappropriate

453

for the recovering of Pd, as it will lead to more complex mixtures and leaching difficulties.

454

Furthermore, the process benefits from previous work, e.g. regeneration of de-soldering

455

baths, and complements existing strategies performed at small industrial scale to recover

456

Au and Cu. Usually, when WEEEs are processed through hydrometallurgy, principal target

457

metal is Au, and other precious metals and base metals are eliminated during a

458

preliminary step in order to maximize Au leaching yield (Figure 7). The proposed process

459

generates a final solid residue which contains still all (or almost all) the Au initially present

460

in the product. This solid residue can be processed according to existing techniques, either

461

directly in the Au leaching step, or with classical base metals dissolution (or even through

462

pyrolysis and grinding for smelting processes). The process also enables selective

463

recovery of Sn as a solid precipitate after dissolution of the solders. We demonstrated that

464

the organic solvent can be re-employed without loss of performance and that the amount

465

of aqueous waste generated is low (ca. 30% of the initial volume to be treated). We have

466

not explored the possibility to recover Ag from this waste, but this option will be explored in

Jo u

rn

al

Pr

e-

pr

oo f

447

18

Journal Pre-proof a near future. Pd is isolated as a Pd(II) salt, dichlorodiammine palladium(II) Pd(NH 3)2Cl2,

468

with an excellent purity (no other trace metals detected). This salt is a key intermediate in

469

Pd chemistry as its burning at ca. 1000°C leads to Pd sponge, basis for Pd fine chemistry.

470

To our knowledge, this study is the first one to detail each step from waste PCBs to a key

471

Pd intermediate at the laboratory scale, including outcome of by-products and effluents.

472

Considering the forecasted increasing demand in Pd, such a process could be further

473

developed and implemented at small industrial scale, i.e. on 10-100 kg of waste PCBs.

474

This is the scale at which regional industrial players operate in the sector of precious

475

metals recovery from waste. These are fully complementary to bigger actors that operate

476

pyrometallurgical processes (Cu smelters). Of course further scale-up of the proposed

477

process is conceivable, and operation of the solvent extraction steps with counter-current

478

continuous

479

investment is higher, and the access to available waste PCBs may be a limit for cost-

480

effective operation rate. Thus we suggest in a first instance not to start at a large industrial

481

scale. Finally, this work, along with others mentioned in this article, strongly suggest

482

players to seek for alternatives to systematic shredding and complete leaching. This

483

approach, inspired by extraction of metals from natural ores, is not adapted to the

484

management of wastes of high content in several metals, of much lower tonnage and

485

diversity.

e-

pr

oo f

467

Jo u

rn

al

Pr

mixer-settlers will lead to better performances. But the technological

486 487

488

Acknowledgements

489

The authors are grateful to the LaBex ChemiSyst for funding the PhD thesis grant of V.

490

Lacanau (ANR-10-LABX-05-01), and to the Commissariat à l’Energie Atomique et aux 19

Journal Pre-proof 491

Energies Alternatives (CEA) for financial support of the study.

Jo u

rn

al

Pr

e-

pr

oo f

492

20

Journal Pre-proof Table 1: Summary of previously published studies dedicated to palladium recovery from waste PCBs. Starting material

Physical treatment

Chemical pretreatment

Crushing, enrichment of metallic components with centrifugal air separator, grinding into metallic particles (94% Cu, 304 mg/kg Pd).

Enrichment through Cu dissolution with CuSO4 and NaCl.

Disasembled components from waste PCBs.

Shredding into metallic particles (18.5% Cu, 97 mg/kg Pd, 86 mg/kg Au).

none

MLCCs.

Grinding, sieving Base metals leaching with concentrated HCl, then with HCl/H2O2.

Disasembled components from waste PCBs.

1

Leaching step CuSO4 + NaCl (Cu/Cu2+ < 0.9) Solution obtained: Pd 20 mg/L, Cu 7260 mg/L + traces Pb, Zn, Fe, Al, Ni.

CuSO4 + NaCl + H2SO4.

l a

o J

n r u

Concentrated HCl/HNO3 at 80°C.

Pd purification

Pd compound obtained

SX with diPd(II) solution isoamyl sulfide, in ammonia stripping with (ca. 100 mg/L) NH3 solution

f o

e

o r p

r P

Not performed

Multimetallic solution (Cu, Fe, Ni, Ag), Pd).

Remarks

Reference

97% Pd recovery during the dissolutionextraction-stripping process, possibility to regenerate the solvent demonstrated.

(Zhang and Zhang, 2014)

58% Pd leaching.

(Yazici and Deveci, 2013)

Au precipitation Not mentioned, Max. 90% Pd with urea and probably leaching. Na2S2O5, then metallic Pd Pd precipitation with Zn dust. Then aqua regia redissolution, and sodium formate precipitation.

(Prabaharan et al., 2016)

Journal Pre-proof MLCCs.

Grinding

None

Concentrated HCl/HNO3 at 25°C. Solution obtained: Pd 108 mg/L, Cu 500 mg/L, Fe 250 mg/L, Pb 700 mg/L, Mn 120 mg/L, Ni 8260 mg/ L, Ba 19 g/L + traces Ag.

Waste PCBs.

Shredding, grinding.

Two steps base metals leaching with H2SO4/H2O2. Then Au leaching with acidic thiourea.

f o

o r p

e

r P

Precipitate of metallic Pd and Ag, Ba and other metals.

Purity depends on SX efficiency: best results given 98.8% Pd purity for 83% Pd recovery yield.

(Fontana et al., 2018)

Regeneration of extraction solvent not evaluated. The authors mention the use of a 10% w/w NaBH4 solution in water, but in these conditions NaBH4 decomposes with release of H2.

2 M HCl + H2O2 Precipitation of Metallic Pd and Safety concerns and NaOCl. both Au and Pd Au precipitate. regarding the with NaBH4. process are not discussed, e.g. Mixing HCl with NaOCl generates Cl2, and adding solid NaBH4 to an aqueous acidic solution leads to a violent reaction generating H2.

n r u

o J

2

l a

SX with quaternary ammonium chloride (Aliquat336) in limonene, then precipitation with NaBH4.

(Behnamfard et al., 2013)

Journal Pre-proof ‘Blast powder’ (industrial waste from a semiconductor processing unit).

None

Waste PCBs.

Shredding and crushing.

Waste PCBs from mobile phones.

None

No leaching

Liquid-liquidpowder extraction with dodecyl amine acetate in kerosene.

Enriched Pd/Ag particles

Supercritical water oxidation, then 0.1 M HCl leaching of base metals.

KI-I2 in acetone and supercritical CO2.

Not performed

Supercritical water oxidation using H2O2, then leaching of base metals with diluted HCl.

Iodine-iodide (KI/I2 at pH = 9)

Composition of powder: metallic Ag/Pd alloy, with Al2O3.

Shredding and grinding.

n r u

l a

o J

3

(Ponou et al., 2018)

Mixture of Ag(I) and Pd(II) iodides in partially iodinated acetone.

Up to 94% extraction of Pd.

(Liu et al., 2016)

Mixture of Ag(I), Au(I) and Pd(II) iodides in water.

Detailed optimisation of process parameters.

f o

o r p

e

r P

The technique focusses on the physical separation between Pd/Ag particles and Al2O3 particles.

Not performed

Concentration of Pd containing solution not given.

Concentration of Pd containing solution not given.

(Xiu et al., 2015)

Journal Pre-proof Table 2: Summary of concentrations (mg/L) of principal metals of interest found during the dissolution/leaching steps of the process (Al, Ba, Nd, Ti, and Y were all found below quantification limit and were removed for clarity).

Ag

Cu

Fe

Ni

Pb

Pd

Sn

Zn

Quantification limit (LoQ)

0,2

0,1

0,2

0,15

0,2

0,3

0,2

0,25

Bath after dissolution of solders


1400

70

180

730


40


After leaching of selected nonmagnetic MLCC

205968

0-4

0-135

5-41

59300

38193

2-24


After dissolution of selected magnetic components

1

2500

2650

3500

700


5


After leaching of components recovered after dissolution of solders

481

1940

350

172

98

33

37

e-

pr

oo f

Composition of solution

Jo u

rn

al

Pr

970

1

Journal Pre-proof Table 3: Distribution coefficients and maximum concentrations expected in organic phase after solvent extraction for unwanted metals present in leaching solution, of which HNO3 concentration is 3 M.

Maximum organic concentration (mg/L)

0,04

7

Cu

0,015

32

Fe

0,24

88

Ni

0,01

2

Pb

0,02

12

Sn

(0,04)a

Zn

0,03

e-

Ag

oo f

Distribution coefficient (D)

pr

Metal

Pr


al

1

Jo u

rn

Note: a. Solubility of tin is very limited in pure nitrate media

2

Journal Pre-proof Table 4: Distribution coefficients of Pd and Fe between a 0,6 M BDMA in toluene organic phase and aqueous phases of variable HNO3 concentrations, and corresponding Pd/Fe selectivity.

Aqueous [HNO 3]

1,2 mol/L

1,5 mol/L

2 mol/L

3 mol/L

D(Pd)

1,7

2,5

3,7

6,1

27

D(Fe)


0,001

0,003

0,024

0,24

SPd/Fe

n.a.

2500

1230

oo f

1 mol/L

250

Jo u

rn

al

Pr

e-

pr

Notes: a. LoQ = 0,001 in the determination of Fe distribution coefficient.

3

110

Journal Pre-proof Table 5: Impact of the presence of chloride anions on the distribution coefficients of key metals present in leaching solution. Conditions: a controlled quantity of NaCl was added to a 3 M aqueous HNO3 phase containing dissolved ions.

D(Pd)

D(Fe)

D(Ag)

D(Sn)

No chloride

27

0,24

0,04

(0,04)a

1 M chloride

0,06

1,0

n.a.b

11,2

oo f

Conditions

Notes: a. Solubility of tin is very limited in pure nitrate media; b. precipitation of silver

Jo u

rn

al

Pr

e-

pr

chloride.

4

Journal Pre-proof Figure 1: Malonamide extractants: BDMA (also referred to as DMDBTDMA), employed in the

study,

and

OEMA

(also

referred

to

as

DMDOHEMA),

Jo u

rn

al

Pr

e-

pr

oo f

actinide(III)/lanthanide(III) separation.

5

employed

in

Journal Pre-proof Figure 2: Details of printed circuit boards (PCBs) employed in the study, and location of palladium in multi-layer ceramic capacitors (MLCC) with electron microscope imaging of a MLCC transversal cut. Pictures of leaching step of the process and solids and solution obtained.

Waste printed circuit boards

Pr

e-

pr

oo f

MLCCs

Pd

al

Dissolution of solders & Recovery of components by filtration

Jo u

rn

Dissolution of components

Leaching solution 80 C

6

4h

Plastics & processors

Journal Pre-proof Figure 3: Results of iron elimination and palladium loss during scrubbing. Initial organic phase based on 0,6 M BDMA in toluene, containing Pd (196 mg/L) and Fe (97 mg/L), scrubbed with different aqueous solutions (H2O or 1,5 M HNO3) at different organic to aqueous phase volume ratios (O/A).

Fe Effet elimination Fe elimination

O/A - lavagePd H2loss O Pertes en Pd

Taux d'élimination du Fe

Effet O/A - lavage HNO3 1,5M Pertes en Pd Fe elimination Pd loss

7,0%

100%

7,0%

90%

6,0%

90%

6,0%

80%

5,0%

80%

5,0%

3,0%

50%

2,0%

40%

1,0% 5

6

7

8

0,0% 9 10 O/A ratioRatio O/A

Jo u

rn

H2O scrubbing

al

30%

7

Pd

pr

60%

70%

4,0%

60%

3,0%

P d

50%

2,0%

40%

1,0%

e-

4,0%

F e

Pr

70%

Fe

oo f

100%

30% 5

6

7

8

9

0,0% 10 O/A ratio Ratio O/A

1,5 M HNO3 scrubbing

Journal Pre-proof Figure 4: Detailed results of the scrubbing sequence optimization. Initial organic phase: 0,6 M BDMA in toluene, containing Pd (95 mg/L) and Fe (79 mg/L), scrubbed with different aqueous solutions (H2O or 1 M HNO3) at different organic to aqueous phase volume ratios (O/A). Two scrubbing stages were performed, at same O/A ratio but with possibly different solutions.

Pd PdPurity purity

Pd loss loss 35%

25%

10%

10 O/A H2O

2nd scrub

H2O / H2O H2O

10

O/A H 2O

H2O / HNO3 1M 1 M HNO3

Jo u

1st scrub

5

rn

5

al

5%

8

92% Série1

Pr

15%

Série4

e-

Pd loss Pd purity

20%

96%

pr

30%

0% O/A ratio

100%

oo f

40%

88%

84% 5

10

1 MO/A HNO3

5

10

1 MO/A HNO3

HNO3 1M / H2O HNO3 1M / HNO3 H2O 1 M HNO3 1M

Journal Pre-proof Figure 5: Flowsheet of the palladium recovery section of the process, along with obtained Pd(II) salt analysis (electron-dispersion analysis and powder X-ray diffractogram). N = number of stages; A/O = aqueous to organic feed ratio; O/A = organic to aqueous feed ratio.

Palladium recovery section

Extraction

Scrubbing

L

L

oo f

Leaching solution

Solvent re-use

N = 2, A/O = 2

N = 2, O/A = 10

HNO3 solution (Cu, Fe…)

L

e-

pr

L

Exp Pd(NH3)2Cl2

Jo u

rn

al

Pr

Ref peaks

9

Stripping N = 1, O/A = 10

Waste solution

L

Pd precipitation S Pd(NH3)2Cl2

L

Journal Pre-proof Figure 6: Detailed results of the solvent-extraction processing section of the palladium recovery process (data for other metals not shown for clarity).

INITIAL AQUEOUS FEED Aq. phase content Pd 98 mg/L Fe 972 mg/L Cu 1942 mg/L

Org. phase content Pd 102 mg/L Fe 184 mg/L Cu 27 mg/L

Extraction yield Pd >99% Fe 19% Cu 1,4%

pr

Aq. phase content Pd < LoQ Fe 795 mg/L Cu 1916 mg/L

oo f

Pd EXTRACTION: 2 stages, cross-current, A/O = 2

Org. phase content Pd 101 mg/L Fe 20 mg/L Cu 2 mg/L

Loss/Extraction yield Pd 1% Fe 89% Cu 93%

Pr

Aq. phase content Pd 16 mg/L Fe 1644 mg/L Cu 250 mg/L

e-

Fe (and Cu) SCRUBBING H2O: O/A = 10

rn

Org. phase content Pd 99 mg/L Fe 3 mg/L Cu < LoQ

Jo u

Aq. phase content Pd 30 mg/L Fe 168 mg/L Cu 5 mg/L

al

Fe SCRUBBING 1 M HNO3: O/A = 10

Loss/Extraction yield Pd 4% Fe 98% Cu >99%

Pd STRIPPING 0,5 M NaCl: O/A = 10 Aq. phase content Pd 854 mg/L Fe 5 mg/L Cu < LoQ

Org. phase content Pd 14 mg/L Fe 2 mg/L Cu < LoQ

Stripping yield Pd 85% Pd purity

99,4%

NEW EXTRACTION CYCLE: 2 stages, cross-current, A/O = 2 Aq. phase content Pd < LoQ Fe 753 mg/L Cu 1922 mg/L

10

Org. phase content Pd 116 mg/L Fe 222 mg/L Cu 23 mg/L

Extraction yield Pd >99% Fe 23% Cu 1,2%

Journal Pre-proof Figure 7: Comparison between existing processes and proposed process: the proposed process dedicated to Pd recovery can fit into existing processes with recycling of Aucontaining solid waste.

Existing hydrometallurgical process

Conc. HNO3

Base metals dissolution

S L

Sieve

Solders dissolution

filtration

Used solution (Pb, Cu…)

Pr

L S

Components dissolution

e-

3M HNO3

Tin residue (Sn)

Jo u

rn

al

Boards (Au, Cu)

11

L S

oo f

Waste solution (Cu, Pd, Sn…)

Proposed process for Pd recovery Prepared waste

Gold leaching

pr

Prepared waste

L S

Au Solid waste (plastics…)

Pd Solid residue (Au, Cu)

Journal Pre-proof REFERENCES

Jo u

rn

al

Pr

e-

pr

oo f

Behnamfard, A., Salarirad, M.M., Veglio, F., 2013. Process development for recovery of copper and precious metals from waste printed circuit boards with emphasize on palladium and gold leaching and precipitation. Waste Manag. 33, 2354–2363. https://doi.org/10.1016/j.wasman.2013.07.017 Birloaga, I., Vegliò, F., 2018. Hydrometallurgical processing of waste printed circuit boards, in: Waste Electrical and Electronic Equipment Recycling. Elsevier, pp. 95–113. https://doi.org/10.1016/B978-0-08-102057-9.00004-4 Choudhary, B.C., Paul, D., Borse, A.U., Garole, D.J., 2017. Recovery of palladium from secondary waste using soluble tannins cross-linked Lagerstroemia speciosa leaves powder: Recovery of palladium from secondary waste using biomass. J. Chem. Technol. Biotechnol. 92, 1667– 1677. https://doi.org/10.1002/jctb.5163 Cole, S.S., 1985. Oxidation and Reduction of Palladium in the Presence of Silver. J. Am. Ceram. Soc. 68, C-106-C–107. https://doi.org/10.1111/j.1151-2916.1985.tb15309.x Cowley, A., 2019. Johnson Matthey Pgm Market Report February 2019. Cui, J., Zhang, L., 2008. Metallurgical recovery of metals from electronic waste: A review. J. Hazard. Mater. 158, 228–256. https://doi.org/10.1016/j.jhazmat.2008.02.001 Diaz, L.A., Lister, T.E., Parkman, J.A., Clark, G.G., 2016. Comprehensive process for the recovery of value and critical materials from electronic waste. J. Clean. Prod. 125, 236–244. https://doi.org/10.1016/j.jclepro.2016.03.061 Ding, Y., Zhang, S., Liu, B., Zheng, H., Chang, C., Ekberg, C., 2019. Recovery of precious metals from electronic waste and spent catalysts: A review. Resour. Conserv. Recycl. 141, 284– 298. https://doi.org/10.1016/j.resconrec.2018.10.041 Facchini, A., Amato, L., Modolo, G., Nannicini, R., Madic, C., Baron, P., 2000. Transient- and Steady-State Concentration Profiles in a DIAMEX-like Countercurrent Process for An(III) + Ln(III) Separation. Sep. Sci. Technol. 35, 1055–1068. https://doi.org/10.1081/SS-100100210 Fontana, D., Pietrantonio, M., Pucciarmati, S., Torelli, G.N., Bonomi, C., Masi, F., 2018. Palladium recovery from monolithic ceramic capacitors by leaching, solvent extraction and reduction. J. Mater. Cycles Waste Manag. 20, 1199–1206. https://doi.org/10.1007/s10163-017-0684-3 Hsu, E., Barmak, K., West, A.C., Park, A.-H.A., 2019. Advancements in the treatment and processing of electronic waste with sustainability: a review of metal extraction and recovery technologies. Green Chem. 21, 919–936. https://doi.org/10.1039/C8GC03688H Işıldar, A., Rene, E.R., van Hullebusch, E.D., Lens, P.N.L., 2018. Electronic waste as a secondary source of critical metals: Management and recovery technologies. Resour. Conserv. Recycl. 135, 296–312. https://doi.org/10.1016/j.resconrec.2017.07.031 Jung, B.H., Park, Y.Y., An, J.W., Kim, S.J., Tran, T., Kim, M.J., 2009. Processing of high purity gold from scraps using diethylene glycol di-N-butyl ether (dibutyl carbitol). Hydrometallurgy 95, 262–266. https://doi.org/10.1016/j.hydromet.2008.06.006 Kaya, M., 2018. Current WEEE recycling solutions, in: Waste Electrical and Electronic Equipment Recycling. Elsevier, pp. 33–93. https://doi.org/10.1016/B978-0-08-102057-9.00003-2 Kolarik, Z., Renard, E.V., 2003. Recovery of value fission platinoids from spent nuclear fuel - Part II: Separation processes. Platin. Met. Rev. 47, 123–131. Le Caër, S., Vigneron, G., Guillaumont, D., Berthon, L., Zorz, N., Moisy, P., 2012. Experimental and theoretical study of the degradation of malonamide extractant molecules under ionizing radiation. RSC Adv. 2, 3954. https://doi.org/10.1039/c2ra01201d Lee, M.-S., Ahn, J.-G., Ahn, J.-W., 2003. Recovery of copper, tin and lead from the spent nitric etching solutions of printed circuit board and regeneration of the etching solution. Hydrometallurgy 70, 23–29. https://doi.org/10.1016/S0304-386X(03)00045-8 12

Journal Pre-proof

Jo u

rn

al

Pr

e-

pr

oo f

Lee, Y.-C., 2010. Dielectric Properties and Reliability of Zn 0.95 Mg 0.05 TiO 3 +0.25TiO 2 MLCCs with Different Pd/Ag Ratios of Electrodes. Int. J. Appl. Ceram. Technol. 7, 71–80. https://doi.org/10.1111/j.1744-7402.2008.02327.x Li, H., Eksteen, J., Oraby, E., 2018. Hydrometallurgical recovery of metals from waste printed circuit boards (WPCBs): Current status and perspectives – A review. Resour. Conserv. Recycl. 139, 122–139. https://doi.org/10.1016/j.resconrec.2018.08.007 Liu, K., Zhang, Z., Zhang, F.-S., 2016. Direct extraction of palladium and silver from waste printed circuit boards powder by supercritical fluids oxidation-extraction process. J. Hazard. Mater. 318, 216–223. https://doi.org/10.1016/j.jhazmat.2016.07.005 Lu, Y., Xu, Z., 2016. Precious metals recovery from waste printed circuit boards: A review for current status and perspective. Resour. Conserv. Recycl. 113, 28–39. https://doi.org/10.1016/j.resconrec.2016.05.007 Mastretta, R., Poirot, R., Bourgeois, D., Meyer, D., 2019. Palladium Isolation and Purification from Nitrate Media: Efficient Process Based on Malonamides. Solvent Extr. Ion Exch. 37, 140– 156. https://doi.org/10.1080/07366299.2019.1630073 Modolo, G., Vijgen, H., Serrano‐ Purroy, D., Christiansen, B., Malmbeck, R., Sorel, C., Baron, P., 2007. DIAMEX Counter‐ Current Extraction Process for Recovery of Trivalent Actinides from Simulated High Active Concentrate. Sep. Sci. Technol. 42, 439–452. https://doi.org/10.1080/01496390601120763 Mojski, M., 1980. Extraction of platinum metals from hydrochloric acid medium with triphenylphosphine solution in 1,2-dichloroethane. Talanta 27, 7–10. https://doi.org/10.1016/0039-9140(80)80003-8 Musikas, C., 1988. Potentiality of Nonorganophosphorus Extractant in Chemical Separations of Actinides. Sep. Sci. Technol. 23, 1211–1226. https://doi.org/10.1080/01496398808075626 Park, Y.J., Fray, D.J., 2009. Recovery of high purity precious metals from printed circuit boards. J. Hazard. Mater. 164, 1152–1158. https://doi.org/10.1016/j.jhazmat.2008.09.043 Petter, P.M.H., Veit, H.M., Bernardes, A.M., 2014. Evaluation of gold and silver leaching from printed circuit board of cellphones. Waste Manag. 34, 475–482. https://doi.org/10.1016/j.wasman.2013.10.032 Poirot, R., Bourgeois, D., Meyer, D., 2014. Palladium Extraction by a Malonamide Derivative (DMDOHEMA) from Nitrate Media: Extraction Behavior and Third Phase Characterization. Solvent Extr. Ion Exch. 32, 529–542. https://doi.org/10.1080/07366299.2014.908587 Poirot, R., Le Goff, X., Diat, O., Bourgeois, D., Meyer, D., 2016. Metal Recognition Driven by Weak Interactions: A Case Study in Solvent Extraction. ChemPhysChem 17, 2112–2117. https://doi.org/10.1002/cphc.201600305 Ponou, J., Wang, L.P., Dodbiba, G., Fujita, T., 2018. Separation of palladium and silver from semiconductor solid waste by means of liquid-liquid-powder extraction using dodecyl amine acetate as a surfactant collector. Sep. Purif. Technol. 191, 86–93. https://doi.org/10.1016/j.seppur.2017.09.009 Prabaharan, G., Barik, S.P., Kumar, B., 2016. A hydrometallurgical process for recovering total metal values from waste monolithic ceramic capacitors. Waste Manag. 52, 302–308. https://doi.org/10.1016/j.wasman.2016.04.010 Serrano-Purroy, D., Baron, P., Christiansen, B., Malmbeck, R., Sorel, C., Glatz, J.-P., 2005. Recovery of minor actinides from HLLW using the DIAMEX process. Radiochim. Acta 93. https://doi.org/10.1524/ract.93.6.351.65642 Sun, Z., Xiao, Y., Agterhuis, H., Sietsma, J., Yang, Y., 2016. Recycling of metals from urban mines – a strategic evaluation. J. Clean. Prod. 112, 2977–2987. https://doi.org/10.1016/j.jclepro.2015.10.116 Sverdrup, H.U., Ragnarsdottir, K.V., 2016. A system dynamics model for platinum group metal supply, market price, depletion of extractable amounts, ore grade, recycling and stocks-in13

Journal Pre-proof

Jo u

rn

al

Pr

e-

pr

oo f

use. Resour. Conserv. Recycl. 114, 130–152. https://doi.org/10.1016/j.resconrec.2016.07.011 Sypula, M., Wilden, A., Schreinemachers, C., Malmbeck, R., Geist, A., Taylor, R., Modolo, G., 2012. Use of Polyaminocarboxylic Acids as Hydrophilic Masking Agents for Fission Products in Actinide Partitioning Processes. Solvent Extr. Ion Exch. 30, 748–764. https://doi.org/10.1080/07366299.2012.700591 Wang, S.F., Dougherty, J.P., Huebner, W., Pepin, J.G., 1994. Silver-Palladium Thick-Film Conductors. J. Am. Ceram. Soc. 77, 3051–3072. https://doi.org/10.1111/j.11512916.1994.tb04549.x Wang, Y., Baker, L.A., Brindle, I.D., 2016. Determination of gold and silver in geological samples by focused infrared digestion: A re-investigation of aqua regia digestion. Talanta 148, 419– 426. https://doi.org/10.1016/j.talanta.2015.11.019 Xiu, F.-R., Qi, Y., Zhang, F.-S., 2015. Leaching of Au, Ag, and Pd from waste printed circuit boards of mobile phone by iodide lixiviant after supercritical water pre-treatment. Waste Manag. 41, 134–141. https://doi.org/10.1016/j.wasman.2015.02.020 Xu, G., Yano, J., Sakai, S., 2019. Recycling Potentials of Precious Metals from End-of-Life Vehicle Parts by Selective Dismantling. Environ. Sci. Technol. 53, 733–742. https://doi.org/10.1021/acs.est.8b04273 Yang, C., Li, J., Tan, Q., Liu, L., Dong, Q., 2017. Green Process of Metal Recycling: Coprocessing Waste Printed Circuit Boards and Spent Tin Stripping Solution. ACS Sustain. Chem. Eng. 5, 3524–3534. https://doi.org/10.1021/acssuschemeng.7b00245 Yang, H., Liu, J., Yang, J., 2011. Leaching copper from shredded particles of waste printed circuit boards. J. Hazard. Mater. 187, 393–400. https://doi.org/10.1016/j.jhazmat.2011.01.051 Yazici, E.Y., Deveci, H., 2013. Extraction of metals from waste printed circuit boards (WPCBs) in H2SO4–CuSO4–NaCl solutions. Hydrometallurgy 139, 30–38. https://doi.org/10.1016/j.hydromet.2013.06.018 Yoo, K., Lee, J., Lee, K., Kim, B., Kim, M., Kim, S., Pandey, B.D., 2012. Recovery of Sn, Ag and Cu from Waste Pb-Free Solder Using Nitric Acid Leaching. Mater. Trans. 53, 2175–2180. https://doi.org/10.2320/matertrans.M2012268 Zhang, X., Guan, J., Guo, Y., Yan, X., Yuan, H., Xu, J., Guo, J., Zhou, Y., Su, R., Guo, Z., 2015. Selective Desoldering Separation of Tin–Lead Alloy for Dismantling of Electronic Components from Printed Circuit Boards. ACS Sustain. Chem. Eng. 3, 1696–1700. https://doi.org/10.1021/acssuschemeng.5b00136 Zhang, Z., Zhang, F.-S., 2014. Selective recovery of palladium from waste printed circuit boards by a novel non-acid process. J. Hazard. Mater. 279, 46–51. https://doi.org/10.1016/j.jhazmat.2014.06.045

14

Journal Pre-proof A simple process for the recovery of palladium from wastes of printed circuit boards Author statement

Jo u

rn

al

Pr

e-

pr

oo f

Damien Bourgeois: Conceptualization, Methodology, Validation, Investigation, Original Draft, Writing - Review & Editing, Project administration Valentin Lacanau: Methodology, Investigation, Writing - Original Draft Régis Mastretta: Investigation Christiane Contino-Pépin: Supervision, Funding acquisition Daniel Meyer: Conceptualization, Supervision, Funding acquisition

15

Writing

-

Journal Pre-proof Declaration of interests X The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Jo u

rn

al

Pr

e-

pr

oo f

☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

16

Journal Pre-proof Highlights A palladium recovery process from waste PCBs was developed at laboratory scale This hydrometallurgical process is a block complementary to existing processes Palladium containing components are separated from the board by de-soldering Leaching with 3 M nitric acid and solvent extraction enable pure palladium isolation

Jo u

rn

al

Pr

e-

pr

oo f

Effluents are well controlled during palladium purification with solvent extraction

17